PROJECT SUMMARY This career development award for Dr. Michael L. Barnett, a primary care physician at Brigham and Women?s Hospital and an assistant professor of health policy and management at the Harvard T.H. School of Public Health, will establish Dr. Barnett as a clinician investigator with expertise on improving the quality and value of care for older adults. This proposal builds on Dr. Barnett?s promising early career researching quality and fragmentation of care delivery and focuses on an increasingly important and understudied domain of fragmented care delivery for older adults: post-acute care (PAC). This K23 award will enable Dr. Barnett to pursue this research area and grow into an independent investigator through support for three training goals: 1) Acquiring skills in advanced quantitative methods for causal inference in large datasets, 2) Developing expertise in PAC for older patients, and 3) Acquiring skills in qualitative research and survey design. Dr. Barnett will pursue these training goals guided by a multidisciplinary faculty team led by primary mentor Dr. Arnold Epstein, an experienced junior faculty mentor with experience performing high-impact research in health care quality and health policy. Dr. Barnett?s co-mentors are Dr. David Grabowski, an expert in post- acute care policy and aging research and Dr. Ateev Mehrotra, who has expertise in episode-based payments and health services research. Dr. Barnett will also partner with distinguished scientific consultants to provide training and mentorship in statistics, geriatrics and qualitative methods. Use of PAC for older adults is growing rapidly, but little is known about whether PAC improves health outcomes for hospitalized older adults. New payment models, such as bundled payments, are increasingly driving providers to lower PAC use, but the potential health impact of this change is unknown. Dr. Barnett will examine the changing landscape of PAC with three integrated aims: first, Dr. Barnett will use Medicare administrative claims to examine the relationship between low rates of PAC use and health outcomes exploiting quasi-random assignment of patients to hospitalist physicians (Aim 1). Next, he will measure the impact on PAC use and health outcomes of a national bundled payment model for total lower joint replacement among high-risk vulnerable older adults who may be negatively impacted by lower PAC use (Aim 2). Finally, he will conduct a comparative series of case studies of hospitals with contrasting changes in PAC use after bundled payment implementation to identify practices associated with adapting to a new payment model (Aim 3). This research can directly inform current policy and payment reform for older adults to enhance the safe and effective use of PAC. Results from this work will form the basis of an R01 proposal to study the relationship between hospital-PAC integration and quality of care for hospitalized older adults.